The fact is, the level of complexity depends entirely on budget.
Most strategy projects are about comparing options, and don't need figures accurate to x decimal places - they need relative figures.
So in deciding between two markets, an approximate modelling process will likely be used (the main inputs, but not worrying about factors that could have a 1%-2% impact but would take a lot of time, and money, to understand).
The main thing to worry about is consistency - selecting inputs that can be obtained for each market / sector and compared side by side.
Of course, sometimes those '1%-2%' factors will be considered, and will make a difference, but clients need to stump up the cash for it. More often than not, they are happy with the cruder option.
I have seen some horribly crude, cringe-worthy models, including some in very big strategy projects where the figures didn't even stand up to the simplest of sense-checks (bravo Deloitte). The sad truth is that crude assumptions are often made and hidden in the (not very deep) depths of these models.